This Briefing Report, prepared in conjunction with the seventh ITU Regulatory Colloquium, concerns the international economic relationships between telecommunications operators which enable them to connect international calls. It focuses on how these relationships are evolving, and on the regulatory policies affecting them. It summarises the existing arrangements concerning accounting rates and settlements; analyses the intense pressures for change; and suggests the kinds of changes that could result.

It also presents the author’s personal view of which outcomes should be preferred, and what choices by national policy makers and regulators will make such outcomes probable. This is strictly a personal view. Neither it, nor any other part of this report, necessarily represents the views of the Colloquium participants, the ITU, or its constituent members: rather, the report represents a starting point for debate on the issues and policy options.

This summary discusses in turn:

• Two radically different kinds of economic relationships for international telecommunications: a new approach based on open competitive markets, with little distinction between international and domestic operations; and the traditional approach based on the exchange of traffic between operators in different countries.
• The existing situation.
• The pressures for change.
• Possible outcomes, summarised in the form of three scenarios.
• The policy issues that governments and regulators face in this field.

1. ECONOMIC RELATIONSHIPS AND PAYMENTS FOR INTERNATIONAL CALLS:
NEW ARRANGEMENTS IN AN OPEN MARKET

The early stages of the transition from monopoly to competition in telecommunications saw a relatively small number of countries move towards a new industry structure for fixed telecommunications based on multiple competing operators. Until the mid-1990s, this had happened in only a few countries: specifically Australia, Chile, Finland, Japan, New Zealand, Sweden, the UK and the US. Since then, governments in some countries have begun to extend to foreign telecommunications operators the rights of competitive entry, and the regulatory protections designed to make competition possible, that they previously offered only to national competitors. For example, within most of the 15-country European Union (EU) such a policy applies to competitive entry by operators from any EU country, with effect from January 1st, 1998. Such changes will transform the marketplace for international telecoms services over the next few years.

Extending pro-competitive policies to foreign telecoms operators has far-reaching implications for the way international calls will be carried, and the associated financial arrangements. Within any group of countries which allow unrestricted market entry by foreign (or foreign-owned) operators and which also grant such operators interconnection rights, there need be little distinction between international calls and those flowing within a single country ("domestic calls"). In such an environment, regulators can and should treat domestic and foreign operators, in a non-discriminatory way, in the interests of fair and efficient competition. This implies that international calls will be carried via the public network in the destination country under arrangements just like those that apply to domestic long-distance calls. International calls will pay interconnect charges that are the same as those for domestic long-distance calls or at least similar. The interconnect service provided is essentially the same in both cases.

For calls within the EU, for example, the difference will disappear as a result of EU legislation. A domestic long-distance call can be brought into, say, Glasgow in Scotland by one of the new competitors to BT can be terminated there via BT’s local network and interconnect service for approximately 0.6 of a UK penny per minute (approximately one US cent at today’s exchange rate). From 1.1.98, the EU legislation gives a French telecoms operator the right to bring a call into Glasgow from Paris and terminate it, again via BT's network, at the same rate.This kind of situation is radically unlike the traditional way of handling international traffic and compensating participating telecoms operators, which we describe in Section 2 of this summary.

International calls will no doubt continue to be priced to end-users somewhat higher than domestic long-distance calls, even for calls that both originate and terminate in countries where competition is unrestricted. The difference is likely to be relatively modest however, reflecting only the (relatively small) difference in network costs; remaining differences in interconnect charges, if any; the greater complexity of business processes for international calls; and probably some persistence of higher-than-average profit margins reflecting the perceived value of international calls.

If the reader objects to this picture on the grounds that it is too idealised and hypothetical, he or she is right on the first count, and wrong on the second. Such situations are not hypothetical: they exist today, and more are coming into existence at a rapid rate. It is true that the description is somewhat idealised: in practice, innumerable regulatory and commercial battles will be fought before the details of the new landscape of a fully competitive international market can be seen clearly. There is, however, no reasonable doubt about the general direction of events.

This report uses the term Single Market to refer to situations of the kind just described. Specifically, this report uses the phrase to mean a group of two or more countries where telecoms operators based in any country in the group can:

• sell international services in any other country in the group.
• extend their network into any other country in the group and establish network Points of Presence (PoPs), there; and further extend its network all the way to the destination customer's premises where that is commercially viable
• interconnect their PoPs to the public switched network of the incumbent telecoms operator in any other country in the group, and terminate international calls (or originate them) on the same interconnect terms and at the same interconnect prices that domestic operators pay.

Single Markets that fit this definition already exist: so far, however, they consist only of a few pairs of countries. The US and the UK form a rough approximation of a Single Market. US operators can and do operate extensively in the UK, largely in the manner described above. To some degree UK-based operators do the same in the US, though they still face severe regulatory obstacles and constraints. End-user prices ("collection rates") for calls between the two countries have plunged. In 1992, BT's "headline" price (before allowing for discount plans) for UK-to-US calls in the peak period was 51 pence (84¢) per minute. By 1995 this price had fallen to around 42 pence (69¢) per minute. By late 1997 BT's headline price had fallen to 24 pence (38¢) per minute; several operators were connecting daytime calls from the UK to the United States at prices in the range 8-15 pence (about 13-25¢) per minute).

Other existing Single Markets comprising pairs of countries include Canada-UK, Sweden-UK, New Zealand-US, and Canada-US. Each of these pairs is characterised by relatively low end-user prices, high and fast-growing traffic volumes, and diverse and fast-changing industry structures and arrangements for handling international traffic.

The domain of Single Markets is about to expand dramatically. It will not be limited to a few pairs of countries but will include large multilateral groups of countries as well. With effect from January 1st, 1998, most of the 15-country European Union (EU) together with the 3 countries associated with the EU through the European Economic Area (EEA), plus Switzerland, will form a Single Market for telecommunications. Telecoms operators from any EU country will be entitled to build network plant and conduct operations anywhere within this Single Market. They will be able to terminate their calls, including international calls, via the incumbent operators' networks. It also seems clear that as a result of the EU Interconnection Directive the interconnect charge payable for international calls will be the same as the interconnect charge for terminating domestic long-distance calls, at least for international traffic within the EU. In practice there will no doubt be the usual delays and difficulties in implementing the new rules. Clearly Single Market conditions in the EU won't come into existence overnight. There are already controversies and delays in adapting national laws, regulations and practices to comply with the EU legislation. During 1997 the European Commission began to put pressure on several EU governments which had been slow in complying with the legislation. After insisting that it would not hesitate to start legal proceedings if they did not stick to the deadlines in the regulatory package, it opened legal "infringement proceedings" against Belgium, Denmark, Germany, Greece, Italy, Luxembourg and Portugal.

An even more extensive Single Market is beginning to emerge. Twenty out of a total of sixty-nine countries participating in the World Trade Organisation (WTO) Basic Telecommunications Agreement of February 15th, 1997 (including nine of the 15 EU countries) agreed to terms which, when fully implemented, will effectively make these countries into a "Single Market" as well. As an informal designation, this report calls these twenty countries the "Group A" countries. A further twelve countries have made commitments which, while somewhat less extensive, still allow considerable scope for trans-national competition.We call these countries "Group B". The two groups are shown together with the criteria defining them in Exhibit ES.1

Exhibit ES.1

THE SINGLE MARKET EMERGING FROM THE 1997 BASIC TELECOMMUNICATIONS AGREEMENT

* Some foreign ownership restrictions on incumbent operators only

The 1997 agreement was originally intended to come into force on January 1st, 1998. This date has been changed to February 5th, 1998 because some of the participating countries had not completed ratification or other national procedures required for confirmation of the agreement on time.

In this wide Single Market, the Basic Telecommunications Agreement requires that telecoms operators based in any of the twenty countries be permitted to operate on the same basis as a domestic operator in any other country among the twenty. Moreover, according to the agreement (following the Most Favoured Nation principle: MFN), the same rights must also be granted to operators from any WTO country. While there are signs that full implementation of this agreement will not be easy, fast or uncontentious, there can be no doubt about the strength or scope of the move towards a Single Market and little doubt that a Single Market will in fact be implemented among most, if not all, of the twenty Group A countries (and possibly among even more WTO countries) within the next few years.

Of course, even after the EU Single Market for telecommunications is in place, and the WTO telecoms agreement is more or less fully implemented, the majority of countries in the world will not yet be members of a Single Market. The great majority of "relations" between pairs of countries for the carriage of international traffic still will not fall within a Single Market, but most of the heavy-traffic routes will be between pairs of countries within a Single Market. Most of the world's international telecoms traffic will flow within one or other of the Single Market groupings, and the Single Market type of environment will consequently shape the way such traffic is carried.

Unrestricted competitive carriage of international traffic in a Single Market is very different from the environment in which the traditional international arrangements were developed. The differences are summed up in Exhibit ES.2. Essentially, the traditional relationships for handling traffic from one country to another (which the ITU calls "jointly provided service") were established between pairs of operators, one in each country. Usually, these operators were monopolies; generally they were authorised only to operate in their own "home" country. The financial arrangements between them which this report refers to as the "traditional arrangements" (and which are described in the next section of the Summary) were well adapted to such a situation, which economists call a bilateral monopoly.

Once the monopoly conditions that gave rise to the traditional arrangements are wholly or partly replaced by a competitive market, challenges to the traditional arrangements, and alternatives to them, naturally emerge. A Single Market opens up many new alternatives to telecoms operators in terms of how they carry their international traffic; what role other operators play in the process; and how (and how much) these other operators are paid.

We refer to these alternatives as "new modes of operation", and describe them below in Section 3 of this Summary. In brief, they are:

• Resale
• Refile, hubbing or re-origination
• International alliances of telecommunications operators
• The extension of foreign operators' networks into the destination countries to Points of Presence (PoPs) they establish there; or in some cases even the extension of foreign operators' networks all the way to customers' premises in those countries. Such arrangements are sometimes called "facilities-based entry" or "self-termination".
• Internet telephony

Contrary to claims often made, the "new modes of operation" will not necessarily entirely replace arrangements of the traditional kind. These may remain in use as one as of the widening range of alternative methods available to operators. Their simplicity means that, in this author’s opinion, they are likely to remain in use for a substantial amount of international traffic, even between "Single Market" countries.

The level of charges (settlement rates) paid by one operator to another within the traditional arrangements is a different matter, however. For the traditional arrangements to remain in use alongside the "new modes of operation", settlement rates will have to fall to levels that are competitive (from the point of view of net payers of settlement payments) with the new alternatives. This means lowering them to the point where paying settlements is comparable to the cost of terminating a call:

• via leased-line resale, or...
• using the foreign operator's trans-national network all the way to a PoP in the destination country, plus interconnect charges between the PoP and the customer, or...
• via the foreign operator’s network all the way to the customer’s premises in the destination country, if the foreign operator has a local network in the distant city (as WorldCom/MFS has in several European cities).

For major operators on major routes between Single Market countries, self-termination may well become the predominant mode of operation.

Most international traffic is still carried today under what this report calls the "traditional arrangements". These are based on international bilateral relationships (referred to as relations) between pairs of nationally-based telecommunications operators, often referred to as "correspondents". Part II of this report describes the details of these arrangements, many (but not all) of which are based on Recommendations of the ITU’s telecommunications standardisation sector, known as ITU-T. The essential features are simple:

• If the number of minutes of traffic flowing from Country "A" to Country "B" exceeds the flow of traffic in the reverse direction, the operator in Country A makes a net payment ("settlement") to the operator in Country B based on the net flow of traffic from A to B (that is, the volume of traffic from A to B, minus the volume from B to A).

• The size of the net payment is calculated from the net flow of traffic (measured in minutes) multiplied by an amount of money per minute agreed upon between the two operators. This amount is based on a negotiated price called the accounting rate. In most cases, the same accounting rate generally applies in both directions: A to B, and B to A. The amount paid per minute from one operator to its "correspondent" is usually half the accounting rate. It is called the settlement rate. (Some national regulators, the FCC in the United States in particular, require the accounting rate to be the same for all operators serving a particular country. Others, like DTI and Oftel in the UK, generally do not.) Modified arrangements apply when the traffic flows from A to B indirectly via a third country (the "transit" country). The accounting rate and settlement rate are intended, according to ITU Recommendations, to reflect the cost of carrying international calls, although how close they are to really doing so is one of the key issues in the current controversy. They have no necessary relationship to the prices charged to end-users for international calls (which are referred to in this context as collection charges or collection rates).

• Where there are multiple operators at one end of the "relation", or at both ends, operating agreements between pairs of correspondents may stipulate proportionate return, though this is not a feature of the ITU Recommendations. Proportionate return means that if an operator in Country A originates, say, 35% of the total traffic flowing from Country A to a particular telecoms operator in Country B, that operator in Country B must send 35% of its return traffic flowing in the other direction back to the same operator in Country A. Regulators in Country A may insist on this, especially if there is a telecoms monopoly in Country B. In particular, a requirement for proportionate return is a major feature of the International Settlements Policy (ISP) of the Federal Communications Commission (FCC) in the United States. On the other hand, regulatory authorities in most other countries with competitive telecoms markets (Australia, for example) do not require proportionate return. Regulators that do require it typically justify this on the grounds that proportionate return prevents the operator in Country B from obtaining excessive bargaining power through allocating return traffic among Country A's several competing operators ("whipsawing"). However, proportionate return tends to favour established operators and raises barriers to entry, and therefore tends to impede the transition to a fully competitive market for international calls.

Those aspects of the "traditional arrangements" that are codified in ITU-T Recommendations have been very influential. While the Recommendations are not legally binding, their main structural features were developed by consensus in the telecommunications industry, and the main features are widely followed, in practice. The key points are reviewed in Chapter 6 of this report. In addition to what we call the "traditional arrangements", the Recommendations do also allow for a variety of alternative arrangements such as "Sender Keeps All" where no settlement payments are made. (In practice, the most important example of SKA is the Internet, which carries international traffic without any form of settlement payment. Its international links, however, which predominantly run to and from the US, are paid for by the non-US Internet Service Providers.)

The "traditional arrangements" have some virtues. They are simple to administer: no minor matter when there are over 200 countries and territories participating in international telecommunications and

thus, in principle, up to 20,000 distinct traffic flows between them that need to be administered. The traditional arrangements have produced a substantial, continuing, and growing international transfer of resources between telecommunications operators. These transfers have helped telecoms operators in some developing countries to expand and upgrade their networks and make progress towards universal service goals. Nevertheless, many critics object to the traditional arrangements on the grounds that they conflict with competitive market principles. Critics often argue that there are alternative and better ways of generating funds to pay for rapid progress towards goals of network expansion and universal service. They also often object that in some developing or newly industrialising countries the flows of hard currency from international settlements are retained by national Finance Ministries and not necessarily used for such investments in the telecoms infrastructure.

The importance of this issue to developing countries (especially those which are relatively small, or have relatively "open", export-oriented economies) should not be under-estimated. In the case of Jamaica, where settlement payments have in most years flowed directly to the telecommunications operator, 46% of telecoms revenue in 1995 was attributable to the total of net settlements received for international traffic. The average for the small island countries of the Caribbean basin is between 30% and 50%. On the other hand, in India total funds received from net settlements though substantial (estimated at US$254 million in 1994), accounted for only 9% the combined total of reported domestic and international telecommunications revenue in 1994, reflecting the large scale of India’s domestic telecoms market.

The impact on developing countries of changes in the international settlement system should be of concern to governments and telecommunications operators in industrialised countries, since (as is often the case) the interests of developed and developing countries are more closely aligned than is generally believed. Notwithstanding their dissatisfaction with today’s level of settlement rates, telecoms operators in advanced industrial countries (and their customers) are beneficiaries of the expansion and upgrading of public networks in developing countries that is to some degree paid for by settlement payments. Telecoms operators in industrialised and developing countries alike almost invariably find that international calls represent one of their most profitable revenue streams.

The traffic that provides the basis of this profitable business is growing rapidly: annual growth of the top five international routes is on the order of 15%, with some routes such as US/Mexico growing by over 20% per year. This does not necessarily mean that revenues are also growing rapidly in every case, because prices are generally falling.   On routes where open competition prevails; they are falling very rapidly, as our UK-US example illustrates.

Future growth prospects are, however, constrained by the low telephone penetration in developing countries and by congestion and other service problems. Congestion results in numerous failed call attempts that incur costs for the telecoms operator trying to originate the call, as well as the operator in the developing country, and generate no revenues. It is very much in the interests of telecoms operators and users (and thus the whole national economy) in advanced industrial countries that operators in developing countries should continue to be able to finance programmes of network expansion and upgrading which are reducing these constraints on worldwide growth in traffic and revenues.

Of course, there are many ways in which the capital investments required by such programmes can be financed. In the case of operators which are privately owned or being privatised, these include foreign direct equity investment; local equity investment; project finance (for example, Build-Operate-Transfer schemes or leveraged leases); and borrowing. But the feasibility of all of these in turn depends on the operator’s profitability, which in many developing countries is closely linked to the flow of net settlement payments.

The shortcomings of the traditional arrangements have become more apparent with the passage of time, as the telecommunications marketplace has become more competitive and more open to entry and to new modes of operation. In particular:

• The structure of these arrangements tends to limit the degree of effective competition on a route. Using the same accounting rate and settlement rate for all operators on a route (sometimes called "parallel accounting") prevents operators competing on price to attract terminating traffic. Proportionate return requirements (which we emphasise are not called for by the ITU Recommendations) obviously also have this effect. Proportionate return puts additional traffic and revenue in the hands of established operators with large outgoing traffic flows, and disadvantages new competitive entrants which might otherwise have been able to bid to carry the terminating traffic. Thus the traditional structure is not very consistent with an overall policy favouring competition.

• Since settlement rates are intended (according to ITU-T Recommendation D140) to be cost-based, and since telecoms operators’ costs per minute of traffic are much higher in some countries (especially developing countries) than others, using the same accounting rate for both directions of traffic seems inappropriate. An "asymmetric" system with different per-minute rates of payment between operators in the two directions would correspond more closely to the economic realities.

• Today’s settlement rates, for calls to a particular country can vary by a ratio of 5:1 or more (sometimes much more) depending on what country the traffic came from. This seems illogical. Settlement rates are intended to be cost based, and yet the charge for terminating an international call in a particular country is largely independent of what other country the call is coming from. The structure of settlement rates should reflect this. This logic has given rise to a proposal for uniform "termination charges" for each country, regardless of the country of origin of the traffic: this approach was, for example, advocated by the Australian delegation to the recent WTO telecommunications negotiations, and is also under intensive consideration in ITU-T Study Group 3.  As opportunities grow for operators to re-route traffic around the world in search of the cheapest alternative routing, some such levelling of settlement rates is inescapable. Technological change and relaxation of regulatory barriers is increasing such re-routing opportunities very rapidly. If levelling of settlement rates does not occur, the routes where higher settlement rates apply will simply be bypassed, a phenomenon akin to "arbitrage" in financial markets. Market forces will ensure an outcome not far from the Australian proposal, though not necessarily in the specific form that was proposed.

• While accounting rates and hence settlements have fallen significantly (for example, settlement rates for traffic to and from the US fell from a weighted average of 51.5 cents per minute in 1992 to 36.5 cents in 1996 and 35 cents in August 1997), it is widely accepted that this fall was insufficient to bring them into line with the true costs of terminating international calls, or to reflect the large cost reductions arising from technological change and economies of scale. The Informal Group of Experts convened by the ITU’s Secretary General to advise on the reform of the settlement system noted in its report of April 1997 that studies carried out by the ITU Secretariat suggested that "in all but a few cases, settlement rates should be priced below 25 cents per minute".

• The structure of the traditional settlement arrangements, combined with the high level of settlement rates, creates economically inefficient incentives for telecoms operators. This structure:

  – limits (though by no means eliminates) the incentive for operators that are net payers of settlements to reduce end-user prices (collection charges) since the settlement rate is an uncontrollable cost representing a substantial percentage of the collection charge.

  – restricts the incentive for operators that are net recipients of settlements to reduce their collection rates and generally promote the growth of outgoing calls through marketing and service improvements (since doing so would increase their outgoing traffic relative to incoming traffic, and thus reduce their net settlement receipts).

• Consequently, the traditional settlement arrangements impede the growth of international traffic, though traffic growth has nevertheless been very rapid.

The incentive structure inherent in the traditional system, to some degree inhibiting the growth of outgoing international traffic from operators that are net recipients of settlements, is one (but only one) cause of the large and fast-growing flows of net settlement payments paid by operators in some industrialised countries.

The United States and Germany have by far the largest net out-payments of settlements, as the "top 10" summary in Exhibit ES.3 shows. In 1995 net payments from the US were US$4.9 billion. The FCC's interim estimate for 1996 is US$5.6 billion. Estimated net payments from Germany in 1994 (the most recent year for which extensive data are available for countries other than the US) were US$0.8 billion; no other net payer country paid more than US$160 million in that year. While settlement payments contribute to the ability of telecoms operators in developing countries to finance their network expansion, the operators paying the settlement payments generally argue that these payments are far larger than what can be economically justified. This is, of course, at the heart of the current controversies about the international payments system. But are the large out-payments for settlements demonstrably inappropriate and harmful, when judged against objective criteria?

It is important to remember that the size of the settlement payments depends on the size of the traffic imbalance as well as the settlement rate. If outgoing and incoming traffic were roughly in balance, the level of the settlement rate would be much less of an issue. The traffic imbalances partly exist for fundamental reasons beyond the control of telecoms operators and regulators: for example, high per capita incomes in advanced industrialised countries; habits of telephone usage in those countries; and the existence of large overseas emigrant communities.

The size and growth of traffic imbalances also reflects other factors that are, to a degree, controllable. Traffic imbalances have grown because competition has reduced collection rates in countries with a competitive telecoms market, despite the constraining effect of settlement payments on reductions in end-user prices. They have also grown because of the phenomenon known as "turnaround". "Country Direct" services such as BT’s "UK Direct" or AT&T "USA Direct" represent one important form of turnaround. If a caller in Country B calls Country A by means of "country direct" services using a telephone credit card/calling card, the call is billed to the customer in Country A, not Country B. The call is counted for settlement purposes as a call from A to B, not from B to A. The same happens for a call from Country B to Country A using a callback operator in Country A. The use of "country direct" service, calling card services or callback in effect increases the measured traffic imbalance that is the starting point for calculating the settlement payments.

The overall effect of the different forms of turnaround, taken together, can be sizeable. In the United States, the ratio of outgoing to incoming international minutes rose from 1.84:1 in 1990 to 2.14:1 in 1995 and 2.29:1 in 1996. In Hong Kong, which has been on the opposite end of this process, the ratio of calls incoming from the US (where a great deal of turnaround activity is based) to calls outgoing to the US was 1.2:1 in 1992. By 1995, this had increased to 3.1, and by the end of 1996 it was 7:1. Since the period 1992-96 is exactly the period during which callback activity, and turnaround in general, burgeoned on this route, it is reasonable to assume that the very large growth in the traffic imbalance is mainly to this.

Such effects, with turnaround leading to increased traffic imbalances, are particularly pronounced for traffic between developed and developing countries. It would plainly be wrong to assume that the traffic imbalances that result in the large and controversial outflows of settlement payments from some industrialised countries are caused solely by the structural defects and inappropriate incentives built into the traditional settlement system. Much of the imbalance is the result of normal competitive market activity, including the creation of service innovations like "country direct" services; though the scale of turnaround activity is of course motivated to a sizeable extent by customers’ desire to avoid the high end-user prices ("collection rates") in monopoly countries.

The mere fact that innovations such as those just described increase traffic imbalances does not mean they are undesirable: in fact, traffic imbalances and resulting flows of settlement payments would not be (or certainly should not be) an issue if the level of the settlement rate represented a fair market price for the service of completing one minute of international calling.

How should the merits and demerits of the traditional settlement system, be judged overall? Any such overall "verdict" must answer four questions:

• Are the economically inefficient incentives built into the traditional arrangements sufficiently harmful to justify making it a goal of public policy to abolish those arrangements?
• Quite apart from the public interest point of view taken by governments, legislators and regulators, is the structure of the traditional arrangements in any case sustainable in the new, increasingly competitive environment?
• Is the level of settlement rates in the traditional system, as distinct from the structure of the system, acceptable from a public policy point of view?
• Is the level of settlement rates sustainable?

This section discusses these questions in turn, giving the author’s personal view.

This report’s answer is "no". The adverse incentives involved in the traditional arrangements, such as discouraging the growth of outgoing traffic from countries that are net recipients of settlement payments, are only economically significant if the settlement rates are above cost-justified levels. But while settlement rates are high, governments and regulators even in countries that are net recipients of settlement payments (mainly developing countries) should be concerned about such disincentives. The way the traditional system constrains the growth of outgoing traffic is bad for the national economies of these countries. Enterprises in these countries should enjoy the same opportunities that companies in other countries have to use telecommunications to explore and pursue their economic opportunities throughout the world. If not, the balance of international competitive advantage in the business environment generally is tilted against telecoms users in the countries that are net recipients of settlements.

The solution is to reduce the settlement rates and not necessarily to abandon the entire structure of the traditional arrangements. Desirable structural reforms include, however, the "levelling" of settlement rates already discussed, and the abolition of proportionate return requirements at least for service to Single Market countries.

Should governments and regulators also allow alternative "modes of operation" such as leased line (private line) resale or foreign-operator PoPs ("self-termination") to grow up alongside the traditional arrangements, or perhaps even direct traffic away from the traditional arrangements? Governments of countries that are net recipients of settlements may be reluctant to allow the use of the new modes of operation, for fear of reducing the settlements payments flowing in. The Federal Communications Commission (FCC) in the United States has also shown itself reluctant in a wide range of situations to allow leased-line resale by foreign operators, and foreign-operator PoPs.

Governments or regulators in other net-payer countries, however have not taken this view: in Japan, New Zealand, Sweden and the UK, for example, use of the "new modes of operation" such as leased-line resale or foreign-operator PoPs is generally permitted regardless of the correspondent country involved. This policy reflects the view, shared by the author of this report, that there is more to gain from the benefits of increased competition than there is to lose from the risk (as yet only hypothetical) that a monopoly correspondent in another country will abuse this liberalisation to gain additional bargaining power and profits.

Regardless of what governments or regulators decide about the merits or demerits of the traditional arrangements, are they sustainable? This report's answer is "yes, but only with significant modification". The reasons are:

Regardless of what governments or regulators decide about the merits or demerits of the traditional arrangements, are they sustainable? This report's answer is "yes, but only with significant modification". The reasons are:

• In a wide variety of situations the traditional arrangements can continue to be an attractive mode of operation for both of the operators making up a "relation". However, this is only the case when the cost of settlement payments (per minute of traffic) to the operator that is the net payer is comparable to the equivalent per-minute cost under new modes of operation such as leased-line resale or foreign-operator PoPs. Sustainability depends mainly on the level of settlement rates, rather than the structure itself.

• Nevertheless, certain modifications are essential for sustainability:

– differences in settlement rates in the same country for different correspondent countries will have to be greatly reduced or even eliminated; otherwise refile and arbitrage will simply divert all the traffic to where the settlement rate is lowest. Anomalies give rise to large arbitrage opportunities for operators other than the traditional operators to capture part of the value of the traffic by re-routing it.

– as a fully competitive environment approaches, it may become impossible to maintain a symmetrical system where the same accounting rate applies to both directions of traffic. The settlement system must provide for lower charges for terminating calls in low-cost countries (this is already provided for within the traditional arrangements in the case of the Europe/Mediterranean region, under the terms of ITU-T Recommendation D300R).

Overall, reform of the structure of the settlement system is a quite distinct question from the level of the settlement rates. There is no logical reason why the broad structure of the existing settlement system (that is, correspondent relations in which operators exchange traffic and a net settlement is paid) should not remain in being if the settlement rates come down sufficiently and the necessary structural changes also take place. The indispensable structural change is the reduction or elimination of differences in settlement rates to the same country based on where the traffic originated; and the use of asymmetric settlement rates where the costs of termination (or origination) of calls are substantially different in the two countries concerned. In this scenario, a modified form of the traditional arrangement continues in use as one mode of operation used for international traffic, coexisting with the various "new modes of operation".

All the complexities of the issue therefore reduce ultimately to three simple questions: when is a settlement rate "too high", and according to what criteria? Who decides that? And how?

The relatively slow rate of reduction of settlement rates, contrasted with the rapid rate of reduction in network costs for international links, has led to a widespread "common sense" view that settlement rates are now too high. Despite the fact that the cost of international transmission links represents only a small part of the total cost of an international call, most independent specialists in this field (including this author) would still support the "common sense" conclusion.

But how, more objectively and analytically, should levels of settlement rates be judged? This is by no means obvious: it is not at all self-evidently valid, for example, to use the text-book microeconomic criteria often employed to argue that settlement rates should equal some measures of a correspondent operator’s Long Run Incremental Costs, as the FCC did in the Notice of Proposed Rulemaking (NPRM) for its International Settlement Rates Proceeding.

It is not realistic to expect a high degree of consensus about this. Not only are there obvious divergences of view between governments, regulators and operators in different countries (especially as between net-payer countries and net recipient countries). Even in theory, it is problematic to ascertain exactly what the "correct" level is. There are good reasons (discussed in detail in Chapter 8 of the main report) to view sceptically even the principle of prescriptions such as the FCC’s proposal for settlement rates based on Total Service Long Run Incremental Cost (TSLRIC), derived from microeconomic theory, let alone the feasibility of applying such prescriptions in practice.

This author’s conclusions are as follows:

• Even from the point of view of economic theory, a pure incremental cost approach (as in the FCC’s assertion in the NPRM that reducing settlement rates to "an incremental cost level... would maximise consumer welfare") is not warranted. Indeed, the FCC itself partially recognised this in its very next sentence, where it added that: "In addition, we believe that it may be appropriate for international services to provide a reasonable contribution to the common costs of foreign carriers".

• In the author’s view, it not only may be, but is, appropriate for settlement rates to provide a contribution to the joint and common costs incurred by an operator to enable its whole network and business to function. Domestic interconnect charges already do just this, in the United States and elsewhere.

• Similarly, since governments and regulators typically impose Universal Service Obligations (USOs) on operators, it is equally appropriate that the international settlement rate should include a contribution to the cost of meeting these obligations: again, this is done in the United States and many other countries in setting domestic interconnect charges. (It is well known, though, that USOs can be an excuse for the costs of monopoly inefficiencies, and that in many areas rural services, operated efficiently, can be profitable and therefore incur no USO costs: any USO contribution should therefore be set at "reasonable", economically justified levels. Agreeing what these are will obviously not be easy, but will be necessary).

• The contribution to joint and common costs and USO costs from international settlements, per minute or as a percentage of the total payment, in general need not necessarily be the same as it is for domestic calls. For routes entirely within a Single Market however, it is not feasible to collect a higher per-minute contribution for international calls than for domestic long-distance calls, since both types of calls can use the same interconnect arrangements

The best that can be hoped for in practice is a series of pragmatic negotiations in which the parties involved eventually find themselves able to agree on:

(a) broadly what elements can legitimately be included in a settlement rate;

(b) what range of numerical values can be arrived at for these elements, under a wide variety of different but logically defensible cost-related doctrines;

(c) what level of settlement rates in any case is sustainable in practice, given the reality that the traditional settlement system is increasingly subject to competitive pressures. Most net recipients of settlements will want to retain the existing system, sufficiently reformed to keep the net payers in the system and thus to achieve sustainability in the face of the emergence of new modes of operation.

The gradual leakage of traffic out of the traditional payment system is steadily compressing the range of settlement rates that will be sustainable in the longer term (say beyond 5 years). It should therefore be possible to get agreement on this pragmatic basis that the great majority of today’s settlement rates are indeed "too high", and to secure faster reductions than have occurred in recent years.

This also seemed to be the view of the ITU Secretary General’s "Informal Expert Group" when it reported in April this year.

No. This does not mean that the more established operators are abandoning the traditional system wholesale, because they must think at least twice before disturbing the existing correspondent relationships upon which so much long-established business relies. But newer players have fewer such concerns. Even many of the traditional players use tools such as refile, or agreements to terminate one-way traffic without a linkage to return traffic, far more often than is generally supposed.

Consequently, traffic is increasingly flowing outside the traditional arrangements as a joint result of the high settlement rates and the evolution towards a new industry structure in which operators have many "new modes of operation" available to them.

Thus there is little practical doubt that today’s levels of settlement rates are not sustainable. Where there is a monopoly operator at one end of a "relation", and "new modes of operation" such as leased-line resale are not allowed at that end, these high levels of settlement rates might in theory be maintained for some years, but even in this case, the arrangement will come under pressure:

• Refile and hubbing will direct traffic to those routes where the settlement rate is lowest.

• If (as usual) collection rates as well as settlement rates are high in the monopoly country, callback and other forms of "turnaround" will tend to inflate the traffic imbalance. This, together with the high settlement rates,may rise the flow of settlement payments to such a high level that the operators (and possibly the regulator) in the net payer country apply very strong pressures for change.

• Major telecommunications users may also press for change, increasingly making unfavourable comparisons between prices and other conditions for telecommunications to the country concerned, compared to other countries.

It follows that substantial reductions in settlement rates are inescapable and that the structure of settlement rates must move towards a uniform charge ("termination charge") for each country, rather than differing settlement rates depending on the country of origin of the call. This will be the case even in most of the countries where a telecoms monopoly still prevails.

In the previous section we discussed the case for change based on fundamental considerations about the desirability and sustainability of international payment arrangements. In this section, we discuss the forces which are likely to challenge the traditional system.

The traditional settlement system and the current level of settlement rates are under challenge from five distinct directions:

• The emergence of "new modes of operation" in which both the operational arrangements and the payment arrangements differ from the traditional correspondent relationship and settlement system.

• Commercial pressures from net payers of settlements. Accounting rates/settlement rates are negotiated between pairs of correspondents, and net payers are becoming more and more reluctant to pay according to today’s (and potentially tomorrow’s) scale of settlement payments.

• Pressure from regulators in net payer countries. Some regulators, notably the FCC, have aligned themselves with the operators based in their countries who are net payers, and are pressing for the reduction of settlement rates. Interestingly, however, regulators in several other countries that are net payers of settlements (notably in Germany, with estimated net settlement out-payments of US$800 million in 1994, and the UK, where 1994 out-payments were estimated at US$158 million) have chosen not to do so.
• The rethinking of international settlement practices that is taking place multilaterally through the ITU and other international forums such as the OECD.

• The WTO Basic Telecoms Agreement. For the thirty-two Group A and Group B countries, whose WTO commitments are in conflict with the traditional settlement system as it is operated today (though not necessarily with a reformed version).

Commercial Negotiations

Commercial negotiations still represent the main avenue for reform in the structure of the settlement system. Notwithstanding the frustrations often expressed in net payer countries, such negotiations (assisted in the case of the US by pressure from the FCC as regulator) have produced sizeable reductions in settlement rates. According to AT&T these "negotiations resulted in a reduction of ...18% in the average accounting rate paid by AT&T from 1992 to 1995. This number is controversial. FCC estimates show larger increases. The FCC has published its estimate (for US international service carriers as a group). It shows a 29% decrease in the weighted-average settlement rate for US operators from 1992 to 1996, and a 32% decrease from 1992 to 1997.However, these reductions certainly still leave settlement rates in most cases well above any level that is likely to have a reasonable cost justification, and for some "relations" there has been little change.

Some major operators, especially AT&T, MCI and others in the US, frustrated with this situation, have been increasingly forceful (and successful) in urging their regulator, the FCC, to intervene.

Increasingly, as telecommunications operators in those industrialised countries that are net payers of settlements have become concerned by the continuing rapid growth and large scale of settlement payments, they have pressed regulators in those countries to become more active in pursuing reform of the settlement systems. Depending on the country concerned, such regulatory activism has included:

(1) Requiring disclosure of accounting rates (which were traditionally kept confidential). This has been done in the US, the UK and New Zealand, and has helped highlight exceptionally high settlement rates and increased the pressure to change them.

(2) In some countries (notably the US), regulators have sought to protect operators in a competitive market against excessive bargaining power at the monopoly end of a "relation" by imposing a mandatory requirement on the operators they regulate to establish proportionate return arrangements with their correspondents in monopoly countries. The FCC in the US does this through its International Settlement Policy (ISP). However, in Australia, New Zealand and Sweden (among other countries) the regulatory rules do not require proportionate return. The rules in the UK do not require proportionate return as such,  but do contain some other safeguards. The FCC also requires uniform accounting rates for all operators to the same correspondent country (a policy sometimes referred to as "parallel accounting"). Here again, regulators in other countries are often less restrictive: UK policy, for example, generally does not require parallel accounting. Policies requiring proportionate return and uniform settlement rates, which arguably could be a safeguard against anti-competitive behaviour by monopoly correspondents, also have their own anti-competitive effects. They raise barriers to entry and favour incumbent operators.

(3) Controlling practices thought to contribute to excessive settlements payments. If there is a competitive market at one end of a "relation" ("Country A"), and a monopoly at the other ("Country B") this may result mainly in the diversion of traffic from B to A (rather than in the reverse direction) out of the traditional settlement system to the "new modes of operation". If this happened, it would further increase the flow of settlements payments in B's favour . Until recently, the UK government for this reason allowed international traffic to be carried via the resale of leased-line capacity ("International Simple Resale" or ISR) only to or from countries that allowed ISR operations by UK operators. This restriction (sometimes referred to as a "reciprocity" policy) was abolished in 1996, through the UK government retained certain safeguards, including regulatory powers to intervene subsequently if anti-competitive behaviour occurs. The Japanese Ministry of Posts and Telecommunications has abolished its ISR restrictions on a similar basis with effect from January 1998. The FCC in the United States used to operate a reciprocity policy (broadly similar to that followed by the UK before 1996), under the name of the "equivalent competitive opportunities" test, or "ECO test". Recently, policies have increasingly diverged. While the UK (like Japan and several other countries) has abolished reciprocity restrictions on resale, the FCC while abolishing the ECO test for traffic between the US and other WTO members as required by the 1997 Basic Telecommunications Agreement has imposed new restrictions (discussed later in this summary) which still greatly restrict leased-line resale

(4) Setting indicative targets for reduction of accounting rates.

(5) Making these targets ("benchmarks") mandatory.

A striking divergence of philosophies seems to be emerging between the FCC on the one hand, and regulators in other countries with competitive telecommunications environments (such as Australia or the UK) on the other. The UK government abolished its ISR restrictions in June 1996, well before the 1997 WTO agreement (which, when it comes into force, will limit national governments’ ability to maintain such restrictions). In abandoning its restrictions, the UK government in effect took the position that gains from an unconditionally open competitive market (expanding the total "size of the pie") outweighed the theoretical risk that such radical deregulation might give too much bargaining power in international telecommunications to monopolies in other countries (enabling them to take too large a "slice" of the expanded pie).

As we describe below the FCC has tended to take the contrary view. It has focused on issues of bargaining power and hence who gets what "slice of the pie". To that end it has maintained strong restrictions on the use of "new modes of operation" for international traffic, except to countries that have already complied with the FCC’s unilateral decisions about how far settlement rates should be reduced. Even for fraffic to or from those countries, significant FCC regulatory constraints and encumbrances still apply. The FCC, through several related but separate regulatory proceedings, has made three key decisions in this field:

• It has indicated that it may waive its International Settlements Policy (ISP), which requires proportionate return and uniform accounting rates for all US operators for any given country. ISP waivers would allow new services, variants of the traditional settlement arrangements, and "new modes of operation" in those cases where it believes this favours the operation of a competitive market. It seems clear, however, that the FCC does not intend to use this flexibility in the sphere of basic fixed telephone services, except in two particular ways:

  (1) to follow a "hands off" policy towards the emergence of Internet telephony.

  (2) to waive restrictions on other "new modes of operation" such as leased-line resale or foreign PoPs (which require FCC approval under Section 214 of the 1934 Communications Act) only under specified conditions linked to the FCC’s policy on settlement rates, as described below.

• It has adopted "benchmark" settlement rates, defined for each of four groups of countries classified by income level. It has in effect given US international operators an ultimatum that they must start applying these benchmark rates by a stated deadline. (The benchmark rates are compared with current settlement rates for a range of countries in Exhibit ES.4.) Since the US operators have operating agreements with their foreign "correspondents", the decision puts them in an interesting position: to comply with the FCC Order, they must renegotiate their settlement rate with the correspondent if they can, and unilaterally breach their agreement if they cannot. Later in this summary (in Section 7 under the heading "Possible Outcomes") we discuss what might happen in practice.

• It has made its consent to the use of several of the "new modes of operation" under which traffic is carried outside the traditional settlement systems (specifically, leased-line resale and foreign-carrier PoPs: see below) conditional to varying degrees on the foreign correspondent carrier complying with the FCC-mandated benchmark settlement rate for the traffic that does continue to be carried under the traditional arrangements. While introducing these new restrictions, the FCC abolished old ones based on a reciprocity principle ("Equal Competitive Opportunities" or ECO test), as required by the commitments to market access and National Treatment made by the US to other WTO countries in the February 1997 WTO agreement.

• It has imposed several new restrictions on the activities of foreign operators using leased-line resale or foreign-operator PoPs in the US; for foreign operators which are considered "dominant" (according to a very all-inclusive definition of "dominant") in their home country.

The FCC's policies undoubtedly represent strong pressure for reducing the settlement rates in the traditional settlement systems. On the other hand, they do not attack the structure of the settlement systems per se. If anything, they tend to do the opposite: they restrict the use of some of the new "modes of operation" through which traffic can be carried outside the traditional settlement systems -- though not, interestingly, the use of Internet telephony.

These FCC decisions raise some major questions about the proper governance of international activities that involve the jurisdiction of both the FCC and regulators in other countries: we discuss the issues involved below in Section 6 of this summary. The decision to link FCC consent for the use of "new modes of operation" (with the exception of Internet telephony) to foreign correspondents’ acceptance of the benchmark settlement rates unilaterally declared by the FCC seems to violate the commitments to market access for foreign operators made by the US in the 1997 WTO telecommunications trade agreement.

We noted earlier how service innovations such as country-direct service or callback result in "turnaround". These particular types of innovations do not take traffic out of the traditional settlement systems: they reverse the effective direction of traffic for purposes of calculating settlements, in such a way as to increase measured traffic imbalances, and flows of settlement payments. Consequently, they also tend to increase the pressures from net payers to curb the growth of settlement payments.

Another distinct set of service innovations, which we refer to generically as the "new modes of operation", have a quite different effect. They either remove traffic from the settlement systems entirely or (as in the case of refile and "hubbing") they involve routing traffic in such a way as to "mix and match" old and new modes of operation to maximum commercial advantage (as an alternative to the term "new modes of operation", the expression "full-circuit regime" has also been used).

The five "new modes of operation" are:

• Resale of leased-line ("private line") capacity to provide public switched international telephone service. In this mode of operation (sometimes called International Simple Resale: ISR) calls originate on the public switched telephone network (PSTN), move to the destination country via leased lines or similar bulk transmission arrangements, and then terminate in the destination country via the PSTN. Freedom to use leased-line resale is incresing rapidly in many countries: in Japan, for example, restrictions on ISR were essentially abolished in January 1998.

• Foreign Points of Presence (PoPs)/Points of Interconnection (PoIs). If an operator from Country A is permitted to extend its own physical network infrastructure, including transmission links, into another Country B and interconnect to the PSTN there in order to terminate international calls, the locations where such interconnection takes place are known as PoPs or (with exactly the same meaning) PoIs. This type of arrangement contrasts with the traditional mode of operation where the operator from Country A possessed only "half circuits" to a notional mid-point between Country A and Country B, not transmission capacity all the way from A to B. This mode of operation is also referred to as "self-termination".

• Refile, hubbing or re-origination, in which an operator takes its international traffic to a country where an open competitive market and low charges apply for forwarding of traffic to its ultimate destination in a "third" country. The traffic may get to the country where this "refiling" occurs either via a conventional correspondent arrangement, via a leased line, or via a foreign-operator PoP. The unconventional routing is selected in order to minimise the originating operator’s cost for terminating international calls. From the point of view of the telecommunications operator where the call terminates, the call appears to have originated from the country where the refiling or hubbing took place: for this reason, refiling or hubbing is sometimes called "anonymous" refile. Such unconventional routings are less and less constrained by technical considerations, because digital signals undergo relatively little impairment even if they traverse very indirect routings: this was not true for the analogue signals that prevailed at the time today’s settlement systems was created. Refile, hubbing and re-origination are quite distinct from the "transit" arrangements that form part of the traditional settlement systems.

• International alliances of operators. Operators may decide to combine their activities in certain lines of business internationally, as in alliances like Concert (led by BT), World Partners (led by AT&T) or Unisource (led by KPN of the Netherlands, Telia of Sweden and Swiss PTT Telecom) that service large multinational business customers. (Proposed transnational mergers take the same trend a stage further.) Such alliances provide end-to-end service. The alliance purchases and pools transmission capacity (either as half circuits or in other forms) provided by the parent company or other telecommunications operators. It uses this capacity to build global networks through which data and value added services, and increasingly voice services as well, are provided.Traffic is not accounted for within the traditional arrangements. The alternative revenue-sharing and cost-sharing arrangements used by the international alliances are complex and diverse.

• Internet Telephony. Recent technological developments, together with the beginnings of gateway arrangements allowing telephone calls to flow between the Internet and the PSTN, opens up a realistic possibility that the carriage of international telephone calls via the Internet ("Internet telephony") will soon move from its original more or less prototype or hobbyist status to become a major "mode of operation" for carrying commercial traffic. It seems so far that this may happen entirely outside the conventional regulatory framework; it is certainly happening outside the traditional settlement systems.

The Rethinking of International Settlements Practices
Through Multilateral Bodies

Just as operators and governments collaborated multilaterally to shape the traditional settlement arrangements, multilateral consultations in recent years have focused on how to adapt international settlement arrangements to the changing environment. Essentially all the players are committed to this multilateral consultative process. Even the FCC, which has chosen to act unilaterally on the question of benchmark settlement rates, has stated that "we should continue to work vigorously with these (multilateral) organisations to pursue accounting rate reform".

The three main multilateral entities active in this field have been:

• the OECD seeking to develop a consensus among governments in advanced market-economy countries.

• Study Group 3 of the ITU’s telecommunications standardisation sector (ITU-T), the source of the ITU’s highly influential formal Recommendations in this field.

• the Informal Expert Group appointed in early 1997 by the ITU’s Secretary General, Dr. Pekka Tarjanne.

The Secretary General himself has played a vital catalytic role as advocate for fresh thinking in this field, pressing the case for rapid reform of the settlement systems along market-oriented lines. Work by the ITU Secretariat, the activities of the OECD, and the ITU-T Study Group 3 have helped clarify the issues and provided useful data and analysis. In addition, the World Bank and various regional bodies including APT in Asia and CITEL in the Americas have been active in stimulating reconsideration of the settlement arrangements. Within the European Union, the European Commission, working with national governments and regulators, has been the architect of the EU Single Market for telecoms services.

The Informal Expert Group (IEG) contributed an independent view on the adapting of international economic relationships and settlement arrangements to an increasingly competitive environment. It puts forward a set of "guiding principles" which:

• favour increased competition and "the move to transparent, non-discriminatory, cost-oriented settlement arrangements";
• advocate "new co-operative relationships" among organisations concerned with the issues, including national regulators, on a multilateral basis;
• emphasise the informational role of the ITU and its contribution to developing costing methodologies and pricing principles;
• propose that the ITU should "help articulate the general range of settlement rates to which current rates are likely to evolve";
• propose a role for the ITU to "mobilise support from other international institutions to help countries make the inevitable adjustments".

In addition, the IEG advocated "an immediate, global reduction in settlement rates of the order of 5 to 10% during 1997, followed by a similar reduction in the first half of 1998". It proposed that "international carriers should guarantee certainty over the transition period for those countries likely to be hardest hit ( e.g. total settlement payments maintained at some predetermined level in exchange for shared risk and staged reductions in the unit settlement rate from current levels to those consistent with effective competition".

Following the IEG recommendation,  ITU in co-operation with the Commonwealth Telecommunications Organisation has since commissioned eight Case Study analyses of the likely impact of reduced settlement revenues on a range of telecoms operators in specific low-income countries.

The 1997 WTO telecoms agreement outlined in Section 1 of this summary will undermine the traditional settlement system, or at least force a radical restructuring and reduction of settlement rates. This will not be the case for all traffic among WTO member countries, but will be the case for traffic originating or terminating in WTO countries that allow the "new modes of operation". As we've explained, allowing the new modes of operation and granting non-discriminatory interconnection rights to foreign operators forces a transition to an open-market environment where charges for originating or terminating international calls via incumbent operators' networks are dramatically reduced.

The vital remaining question, therefore, is just which traffic flows regulators will allow the telecoms operators to interconnect in this way. This is a complex and difficult question, turning on how far some national regulators will try to maintain restrictions, and how far the treaty obligations in the GATS and the 1997 agreement turn out to permit or prevent such restrictions. Over the three or four years, however, we believe the effect of the WTO agreement will be to allow the new modes of operation and thus revolutionise the arrangements for international traffic along Single Market lines, at least for traffic among most (perhaps all) of the Group A countries.

Considering what might happen next is inevitably speculative in such a fast-changing situation. Still, it’s useful to consider in broad terms the kinds of changes that may come about, and also how they may come about.

There are essentially four ways that the international economic relationships involved in the carriage of international traffic may change:

(1) The traditional arrangements might remain in use for a large amount of international traffic, but with the level of settlement rates, at least for the great majority of "relations", being substantially reduced.

(2) The traditional arrangements may be restructured. This might include cost-based "termination charges" which would be the same for all international calls to a given country, regardless of the country where the call originates; or at least would involve settlement rates which vary relatively little for each destination country. It might also include asymmetric settlement rates which are lower for high-volume, low cost operators in advanced industrialised countries than for operators with lower volume and higher costs, predominantly in developing countries.

(3) Transferring traffic to "new modes of operation" outside the settlement systems.

(4)  The distinction between international service and domestic long distance may be abolished, so that domestic interconnect charges replace settlement payments.

In practice, it is unlikely that any one of these four possibilities will prevail everywhere. There will be a mix, different for different pairs of countries:

• Options (1) or (2) may predominate for traffic to and from many developing countries, although it is also likely that an increasing number of developing countries may make a radical step towards an open-market, pro-competition policy by joining a "Single Market", in which case options (3) or (4) also come into play. Option (2) is currently the focus of Intense consultative activity in ITU-T Study Group 3.

• For traffic between countries that have pro-competitive regulatory rules, to varying degrees, the various "new modes of operation" (Option 3) are likely to become more and important: large amounts of traffic will be handled in those ways outside the framework of the traditional settlement arrangements.

• For pairs or groups of such countries already forming, or committed to form a Single Market (US/UK for example, or EU/EEA), it is already clear that option (4) or at least something very close to it, will eventually predominate. This does not necessarily mean, however, that the traditional settlement arrangements will vanish for all international traffic within a Single Market. Operators will sometimes find it convenient managerially and perhaps also the most economic solution, especially for lower-volume traffic flows, to continue using the traditional system, as long as they can do so at settlement rates which are not far above the level of domestic interconnect charges.

6. HANDLING THE TRANSITION:
ISSUES OF POLICY AND GOVERNANCE

Policy Choices: Regulatory Options and Governance

The developments and issues described in this report require government decision-makers (Ministers, executive officials, legislators, regulators, and even courts of law, depending on the structure of government in each particular country) to make choices about two distinct kinds of questions: process questions concerning how regulatory decisions should be made and, if necessary, enforced; and substantive questions about the regulatory decisions themselves.

"Process" Questions

National decision-makers will increasingly need to choose...:

• How far to leave settlement issues to be decided by the telecommunications operators through commercial negotiations, and how far to intervene.
• If the decision is to intervene, whether to act:

– Unilaterally, telling telecommunications operators carrying traffic to or from particular countries what they must do or may not do concerning settlement payments, without reference to the views of the regulator (or other relevant government authority) in the other country concerned.

– Bilaterally, through decisions concerning individual "relations" with particular correspondent countries, arrived at through agreements or understandings with each of those countries’ National Regulatory Authorities (NRAs).

– Multilaterally, through agreements or understandings (e.g. voluntary compliance with new or revised ITU-T Recommendations) reached among NRAs and operators from numerous countries, through the ITU or otherwise.

On the first question, the stances of each of the NRAs in different countries fall fairly clearly into three categories:

• Interventionist: actively intervening, and specifically seeking to impose levels of settlement rates well below today’s levels. So far the FCC in the United States is the only regulator in this category.

• Pro-competitive but non-interventionist: NRAs in most of the countries with a pro-competitive telecommunications policy are in this category. They are aware of the issues concerning settlements and have adopted some regulatory measures in this field (usually concerning disclosure), but have chosen not to seek to impose a solution on the operators or on regulators in other countries. It is striking that in several countries other than the US that now have pro-competitive regulatory regimes and are net payers of settlements including Australia, Germany, Hong Kong and the UK, the national regulators have so far (despite active consideration of the issues) chosen not to intervene.

• Regulators in countries with a monopoly industry structure for telecommunications: So far these NRAs, where they exist, have either not addressed the subject of international settlements at all, or have favoured the status quo.

Until recently, settlement matters were handled almost exclusively by telecoms operators themselves. The operators defined the broad framework of the settlement systems by creating a consensus for particular ITU-T Recommendations. They negotiated the specifics, such as levels of accounting rates, through commercial negotiations between themselves, separately for each pair of operators (i.e. each "relation"). As the US, followed by other countries, made the transition from monopoly to competition in telecommunications, the pressures for change began to escalate. So far, however, the FCC remains the only NRA in the world to have taken a highly activist stance about the settlement system. Some other NRAs have taken cautious steps in such areas as transparency (e.g. Oftel in the UK has published all UK accounting rates) and have retained reserve powers to intervene in future if necessary, especially to ensure that new operators are fairly treated. None has yet undertaken a radical intervention comparable to the FCC’s benchmarks decision of August 1997.

On the matter of "unilateral", "bilateral", or "multilateral" action, two essential questions are inescapable:

• Does an NRA’s jurisdiction extend to unilaterally deciding the terms of an international collaboration between telecoms operators in different countries to provide international service? (In such a collaboration, each of the participating telecoms operators should be a "willing buyer" and "willing seller" of services, and the sovereignity of the correspondent country and the jurisdiction of its NRA should be respected.)

• Even if the answer to the first question is "yes", is it wise to seek to dictate such decisions unilaterally?

The answer to the first of these two questions depends on complex issues of both domestic and international law. It is clearly beyond the scope of this report, and its author’s expertise, to give a view on the legal issues involved. It is however relevant to note that, over the years, a body of legal principles and precedents has grown up concerning the handling of international commercial and operational issues that involve two or more national jurisdictions. One of the key principles is that of "comity": the obligation of courts and other public bodies in one country to give due weight to the jurisdiction and laws of the other country or countries involved.

The FCC, in taking the unilateral route, stated (in its "Benchmark" Report and Order of August 1997) that:

• "...we will require that US carriers negotiate with their foreign correspondents settlement rates at or below the appropriate benchmark... If US carriers fail to achieve progress... we will take... enforcement action..."

• it will only grant foreign or foreign-affiliated operator authority to establish PoPs in the United States ("certain types of Section 214 authorisations") on condition that the "foreign affiliate offer US international carriers a settlement rate at or below the relevant benchmark".

Startlingly, the FCC claims that the first of these provisions (among others) does "not constitute an exercise of jurisdiction over foreign carriers" since the decision and any related enforcement actions "will apply to US carriers within our jurisdiction, not their foreign correspondents", and the decisions will only have an "indirect" effect on operators outside the US. It also argues that its position on the granting of Section 214 authorisations does not conflict with the commitments of the United States to market access and National Treatment of foreign operators under the 1997 WTO agreement, even though the US Schedule to that agreement says nothing about the commitments being conditional upon reduced settlement rates.The General Agreement on Trade in Services (GATS) and the 1997 agreement allow for measures by national regulators to combat anti-competitive behaviour by companies with market power. The GATS requires, however, that these measures must not be such as to "nullify or impair" the country's commitments to open their national markets to competition.

No doubt these issues will be very fully dissected in the US District Court for the District of Columbia. As of October 23rd, 1997, six non-US telecoms operators and two international bodies had submitted "Petitions to Review" the FCC’s August decisions in the "Benchmark" proceeding, typically arguing that (to quote Cable & Wireless plc of the UK), these decisions establish "without... jurisdictional authority the rate that foreign common carriers... must charge US common carriers for terminating their traffic in the foreign market...". Separately, the Philippine regulator (the National Telecommunications Commission), and the largest Philippine operating company (the Philippines Long Distance Telephone Company) have filed "Petitions for Reconsideration" with the FCC.

But aside from the question of the lawfulness of initiatives to impose an outcome unilaterally, there is also the question of whether such an approach is wise. The entire fabric of international telecommunications has been built up based on a very high degree of mutual and voluntary co-operation between operators and governments in widely varying societies around the world. This co-operative process has become successful beyond the wildest dreams of its founders in fields ranging from numbering and standards for technical and operational compatibility. Any user of computers can testify that such seamless connectivity and universal compatibility was not guaranteed to happen, does not in fact happen in large areas of computer applications, and should not be taken for granted. Unilateral action on key issues does not enhance the atmosphere for such voluntary co-operation in the future.

To the extent that national regulators (individually, or collectively through a multilateral process) do decide to intervene in the international settlement systems (as the FCC clearly has decided to do, but other NRAs apparently have not), the agenda for the regulator in this field includes:

• What levels of settlement rates to insist upon, based on what economic or regulatory principles.

• Whether to regulate the non-price terms of operating agreements, e.g. requiring uniform settlement rates for all operators serving a particular country, or requiring proportionate return.

• Whether to attempt to restrict the "new modes of operation", tolerate them, or positively encourage them.

In developing countries, the agenda also should include:

• Steps to ensure that operators exploit the advantages to them of the "new modes of operation" (e.g. foreign-operator PoPs) to generate additional outbound traffic to industrialised countries, outside the settlement systems, wherever regulatory rules and operational practicalities permit. Such measures could, for example, include bilateral agreements with industrialised countries, or negotiated entry into multilateral Single Market groupings for purposes such as that emerging from the 1997 WTO agreement.

• Measures to minimise the adverse effect of reduced settlement in-payments on telecommunications network development and on progress towards universal service goals: for example rebalancing of prices to make activities other than inbound international traffic more profitable.

Even where these possibilities are energetically pursued, however, it is likely that the forthcoming reductions in settlement rates will cause significant disruptions to the financial management of many telecoms operators in developing countries.

Stepping back from the intricacy of the issues and possibilities discussed so far, is it possible to discern broadly how events might develop? This section suggests a range of possibilities, in the form of three scenarios.

Scenario 1: The "Soft Landing" Scenario

This scenario broadly corresponds to the changes envisaged by the Informal Experts Group (IEG) appointed by the Secretary General of the ITU, which reported in April, 1997. It involves a degree of "give and take" on the part of all the players involved:

• For high-income countries, which also are becoming (with few exceptions) countries with open competitive telecoms markets as well, changes that are naturally taking place as a result of the transition to a Single Market lead to rapid reductions in settlement rates. This ensures that there is no significant or lasting conflict with the FCC's "benchmarks". In any case, for traffic between the various Single Market countries, large amounts of traffic flowing via the "new modes of operation" begin to be terminated at domestic interconnect rates, at termination rates close to domestic interconnect rates, or at end-user prices. These are far below the benchmarks, and settlement rates are commercially negotiated down to the benchmark levels or below in order to ensure sustainability.

• For other countries, including major net recipients of settlement, negotiations lead to staged but accelerated reductions in settlement rates, but not in all cases rather than the very rapid reductions called for by the FCC’s benchmarks and associated deadlines.

• Developing countries which are adversely affected are able to obtain significant transitional assistance from the World Bank, the ITU and other international agencies.

Scenario 2: The Conflict Scenario

In this scenario, the "irresistible force meets the immovable object". Net recipients of settlements refuse to lower settlement rates fast enough to satisfy the demands of the telecommunications operators who are large net payers (AT&T, for example), and the FCC. Jurisdictional issues escalate (starting from the District Court proceedings that began in September 1997), as the FCC and the major net payers based in the US seek to force the issue.

The likely outcome would be a protracted stalemate while the jurisdictional issues are fought out in the courts or through other dispute-resolution mechanisms, possibly even high-level international diplomacy.

As this scenario proceeds (if it does), international relationships between operators, governments and regulators would be likely to deteriorate, possibly culminating in a variety of adversarial acts, which might include:

• One or more operators in a net-payer country unilaterally reducing settlement payments.

• A regulator in a net-payer country (probably the FCC) ordering a carrier to do this.

• An operator in the correspondent country, either spontaneously or under the direction of its regulator, inactivating certain international circuits in retaliation, or at least refusing to authorise additional circuits; or alternatively sending all its return traffic indirectly via an operator in a third country.

• The operator in the net payer country routing all its traffic to the net-payer country by indirect routings.

Lest this scenario seem alarmist, it is worth noting that events along exactly these lines occurred in the case of the "relation" between AT&T in the US and Telintar in Argentina in 1996, though the

dispute apparently has been settled subsequently. Past experiences with previous attempts to unilaterally decide international issues are also not encouraging as precedents in the orderly management of international telecommunications.

The case for maximum effort to avoid the Conflict Scenario is compelling.

Scenario 3: The Competitive Response Scenario

The essential distinguishing feature of this Scenario is that in it, countries which are currently major net recipients of settlements, especially developing countries, would positively embrace the opportunities offered by a competitive international environment. They would seek to maximise these opportunities rather than simply accepting reduced settlement rates and seeking to minimise their adverse impacts (as in Scenario 1), or seeking to delay change and defend the status quo (as in Scenario 2). In doing so, an essential element of the strategy of developing countries would be to hold governments and regulators in the industrialised "net payer" countries fully to their own expressed commitments to open competition.

In this scenario, as in Scenario 1, operators who are net recipients of settlements today agree to accelerated reductions of settlement rates, but do so on condition that they are granted the right to carry traffic into the net-payer countries such as the UK or the US using "new modes of operation", with no further regulatory obstacles or encumbrances. Such initiatives could involve Internet telephony, leased line resale or their own PoPs established in those countries. Of course, relatively few operators in developing countries will have the resources to implement such a strategy on their own. It is nevertheless entirely practical to do so through an alliance or through an arrangement in which a strong operator from one developing country carries traffic for other developing country operators. Alternatively, some third party, which might well be a telecoms operator based in an industrialised country, could perform the necessary aggregation of traffic and operate the necessary network arrangements under contract to several developing-country operators.

In effect, governments and telecoms operators from developing countries would (in the US case) be turning around the FCC’s position of refusing to authorise new modes of operation such as leased-line resale or foreign-operator PoPs until the correspondent country’s settlement rate is reduced to (or below) the benchmark rate promulgated by the FCC. The correspondent would pursue the converse proposition: agreeing to accelerated reductions in settlement rates, on condition that it receives the right to utilise "new modes of operation": in the US, for example, it would petition the FCC to grant Section 214 authority, and full interconnection rights, for such an extension of the correspondent’s operations into the United States (for example via a foreign-operator PoP), with no further conditions or restrictions other than the reduction of settlement rates to the agreed levels. For correspondents from countries that are WTO members, this is a matter of holding the industrialised, net-payer countries to the letter of their market-opening commitments in the 1997 WTO telecommunications agreement.

In the new Single Market situation thus created, the operator from the developing country would compensate for its reduced settlement rate from incoming international calls by a large expansion of its profits on two new other lines of business. The first of these is international traffic to each major country that is a net payer of settlements, carried outside the traditional settlement system and using one or more of the "new modes of operation". The second is carriage of traffic from such countries under turnaround arrangements such as card calling or "country direct" services. The increased profits would arise from:

• Increased volume, due to reduced collection rates and increased marketing and product/service innovations and enhancements, for which there would now be a much greater incentive. (Provision of card-calling and "country direct" services to overseas emigrant communities is an important example.)

• Low unit costs, due to the use of end-to-end transmission links and domestic interconnect at the distant end of the call in the "net payer" country.

To fully exploit the Scenario 3 opportunity in practice, groups of developing countries would need to find ways to aggregate their outgoing traffic to achieve economies of scale. The scenario indicated several ways this can be done. Likely candidates to launch such an initiative are operators that have emerged or are emerging as major "hubbing" players in developing or newly-industrialised countries: for example, Singapore Telecom arguably is such a player already, and VSNL in India apparently intends to become one.

Contemplating the relatively unpromising prospects offered for some developing countries by the "Soft Landing" scenario, and the unconstructive nature of the Conflict Scenario, it seems worthwhile to seriously investigate the possibilities of Scenario 3, the Competitive Response Scenario. Above all, every effort should be made to avert the Conflict Scenario. At worst, it might disrupt the orderly management of international services, or conceivably even interrupt service between certain countries at certain times. Even on the most optimistic view, it would surely undermine the excellent working relationships that have so far made international telecommunications work so well.

The best opportunity for developing countries is not to concentrate exclusively on seeking to minimise the damage caused by reductions in incoming settlements (though this is obviously necessary), but to re-examine the entire configuration of their business to obtain the best economic results achievable in the new environment. In particular, operators in developing countries should:

• Respond to the increased incentive to cut collection rates and expanding outgoing traffic, through a major effort to:

- Increase outgoing volume not only by greatly reducing collection rates but also by extending and enhancing services; by improved marketing and service innovations (for example, through innovations such as pre-paid telephone cards, and other kinds of card services); and by attracting private capital into what can certainly be very profitable investments in expanded international operations andinternational capacity.

- Increase margins for outgoing traffic through aggregation of traffic from multiple countries; acquiring end-to-end transmission capacity instead of half circuits; and establishing PoPs in industrialised countries. This of course must involve governments in a major effort to negotiate away remaining regulatory barriers.

- Make full use of the "new modes of operation", including Internet telephony.

- Use reverse call-back from developing countries to industrialised countries where appropriate.

• Develop a negotiating approach for bargaining with regulators and operators in the industrialised countries to ensure that regulatory barriers in those countries which could block implementation of such a strategy are removed. A key issue is the acceptance of asymmetrical interconnection or termination charge, in which the overall level of charge is reduced but the legitimate cost-based case for setting such rates higher in developing countries than in industrialised countries is recognized.

• Recognise that achieving this in practice is likely to require concessions to the concerns of industrialised countries where the telecoms market is competitive. Such concessions might include agreement to reduced settlement rates matching, or reasonably close to, what is being sought by the FCC, or new commitments (through the WTO process, for example) to a phased timetable for increased opening of the national market to international competition.

• Develop strong "country direct", card calling and callback services of their own. These were less attractive in the past because they would have reduced the incoming traffic imbalance and hence the level of settlement payments, but become more and more attractive as settlement rates within the traditional correspondent system move down towards cost.

The new environment, while highly challenging for almost every participant, need not be a "zero sum game". Gains for one participant do not necessarily mean setbacks for another. The key to managing the new situation successfully is to recognise this fully, and seek out so-called "win-win" solutions where as many participants as possible can achieve a positive outcome.